Session: 07-10-02: Medical Robotics, Rehabilitation, and Surgery

Paper Number: 191309

Design and Implementation of an Integrated Kalman-BMFLC Algorithm for Physiological Tremor Suppression in Surgical Robot 

This paper focuses on involuntary physiological tremor, a periodic phenomenon inherent in healthy human subjects. In high-precision teleoperated surgical robot systems, tremor signals during master manipulator operations degrade surgical precision and increase operative risks. Current research on physiological tremor characteristics remains insufficient, with most studies concentrated in the medical field, lacking accurate mathematical modeling of physiological tremor signals. This paper investigates tremor suppression algorithms for physiological tremor. Pure tremor signals were acquired using an inertial measurement unit. Fast Fourier Transform (FFT) frequency-domain analysis of collected physiological tremor data from healthy subjects confirms that tremor frequencies are concentrated within the 8–14 Hz band. This frequency band is consistent with the typical frequency range of human physiological tremor and exhibits good separability from intentional motion frequencies in surgical operations (usually below 2 Hz), providing a theoretical basis for the design of subsequent filtering algorithms. Kalman filtering is an optimal linear recursive estimation algorithm based on state-space models, capable of achieving optimal state estimation under the minimum mean square error criterion. The Band-Limited Multiple Fourier Linear Combiner (BMFLC) is an adaptive algorithm specifically developed for physiological tremor signal filtering, which employs a set of Fourier series covering a specific frequency band within a linear combiner to accurately estimate and suppress multi-frequency tremor signals within this band. Based on the frequency-domain characteristics of tremor, this study investigates the application of Kalman algorithm and BMFLC for physiological tremor　suppression. However, Kalman filtering alone fails to achieve accurate separation of intentional motion and tremor signals, while BMFLC—though capable of adaptive tremor estimation—exhibits constrained convergence speed and steady-state precision inherent to its fixed-step-size adaptive update scheme. To address these limitations, this paper proposes a joint estimation algorithm combining Kalman filtering with BMFLC (JK-BMFLC). This algorithm unifies intentional motion states and BMFLC tremor parameter weights within a single state space for joint optimal estimation via the Kalman algorithm. The core innovation of this algorithm lies in utilizing the Kalman filtering framework to optimally estimate BMFLC weight coefficients, rather than traditional gradient descent updates, thereby improving parameter convergence speed and estimation accuracy. To validate algorithm performance, simulated tremor signals were established firstly by superimposing low-frequency intentional motion, 8–14 Hz physiological tremor, and Gaussian white noise, capable of faithfully reflecting signal characteristics of master hand operations. The suppression performance of Kalman, BMFLC, and JK-BMFLC algorithms was analyzed and compared, with results demonstrating that the JK-BMFLC algorithm reduces root mean square error by 40%. Finally, the feasibility of the joint estimation algorithm for physiological tremor suppression in oral surgical robots was validated using operational signals collected from professional surgeons on the oral surgical robot system. Experimental results indicate that the JK-BMFLC algorithm can effectively suppress physiological tremor in surgeon operations while maintaining real-time tracking of intentional motion commands, significantly improving the operational precision and stability of surgical robots, possessing important clinical application value.

Presenting Author: Zhile Wu Hefei University of Technology

Presenting Author Biography: Zhile Wu received the B.S. degree in engineering from South China Agricultural University, China, in 2021, and the M.S. degree in mechanical engineering from Hefei University of Technology in 2024. He is currently a Ph.D. candidate at the School of Mechanical Engineering, Hefei University of Technology. His research focuses on surgical robotics and physiological tremor suppression algorithms.

Authors:

Zhile Wu Hefei University of Technology
Shuo Jiang Hefei University of Technology
Hao Zhang Hefei University of Technology
Yuan Li Hefei University of Technology
Ping Zhao Hefei University of Technology